{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        os.path.join(dirname, filename)\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2023-06-10T21:34:37.429453Z","iopub.execute_input":"2023-06-10T21:34:37.429830Z","iopub.status.idle":"2023-06-10T21:34:53.330249Z","shell.execute_reply.started":"2023-06-10T21:34:37.429798Z","shell.execute_reply":"2023-06-10T21:34:53.329355Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import os, shutil, pathlib\nimport pandas as pd\nfrom sklearn.model_selection import train_test_split\nfrom tensorflow.keras.preprocessing.image import ImageDataGenerator\nimport numpy as np\nfrom tensorflow import keras\nfrom tensorflow.keras import layers\nimport glob\nfrom sklearn.base import BaseEstimator, ClassifierMixin, TransformerMixin\nfrom sklearn.pipeline import Pipeline\nfrom sklearn.metrics import accuracy_score","metadata":{"execution":{"iopub.status.busy":"2023-06-10T21:34:53.338602Z","iopub.execute_input":"2023-06-10T21:34:53.338885Z","iopub.status.idle":"2023-06-10T21:34:53.353366Z","shell.execute_reply.started":"2023-06-10T21:34:53.338861Z","shell.execute_reply":"2023-06-10T21:34:53.352406Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"data_dir = \"/kaggle/input/dog-breed-identification\"\ntrain_dir = os.path.join(data_dir, \"train\")\ntest_dir = os.path.join(data_dir, \"test\")\nlabels_df = pd.read_csv(os.path.join(data_dir, \"labels.csv\"))","metadata":{"execution":{"iopub.status.busy":"2023-06-10T21:34:53.357422Z","iopub.execute_input":"2023-06-10T21:34:53.359294Z","iopub.status.idle":"2023-06-10T21:34:53.375133Z","shell.execute_reply.started":"2023-06-10T21:34:53.359262Z","shell.execute_reply":"2023-06-10T21:34:53.374375Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"labels_df.head()","metadata":{"execution":{"iopub.status.busy":"2023-06-10T21:34:11.101655Z","iopub.execute_input":"2023-06-10T21:34:11.102034Z","iopub.status.idle":"2023-06-10T21:34:11.111113Z","shell.execute_reply.started":"2023-06-10T21:34:11.102007Z","shell.execute_reply":"2023-06-10T21:34:11.110119Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"x_train = []\ny_train = []\nx_test = []\n\n# Obter a lista de arquivos .jpg no diretório de teste\njpg_files = glob.glob(os.path.join(test_dir, \"*.jpg\"))\n\n# Carregar todas as imagens de uma vez\nfor jpg_file in jpg_files:\n    image = keras.utils.load_img(path=jpg_file, target_size=(224, 224))\n    input_arr = keras.utils.img_to_array(image)\n    x_test.append(input_arr)\n    _id = filename.split(\".\")[0]\n\n\n# Converter a lista x_test em um array numpy\nx_test = np.stack(x_test)\n\n# Obter a lista de arquivos .jpg no diretório de teste\njpg_files = glob.glob(os.path.join(train_dir, \"*.jpg\"))\n\n# Carregar todas as imagens de uma vez\nfor jpg_file in jpg_files:\n    image = keras.utils.load_img(path=jpg_file, target_size=(224, 224))\n    input_arr = keras.utils.img_to_array(image)\n    input_arr = np.array([input_arr]) \n    _id = filename.split(\".\")[0] \n    ticket = labels_df[labels_df[\"id\"] == _id][\"breed\"].values[0]\n    x_train.append(input_arr)\n    y_train.append(ticket)","metadata":{"execution":{"iopub.status.busy":"2023-06-10T21:34:56.686530Z","iopub.execute_input":"2023-06-10T21:34:56.686899Z","iopub.status.idle":"2023-06-10T21:38:19.701330Z","shell.execute_reply.started":"2023-06-10T21:34:56.686870Z","shell.execute_reply":"2023-06-10T21:38:19.700317Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"x_train_splited = x_train[:5000]\ny_train_splited= y_train[:5000]\n\ny_train_to_test = y_train[5001:]\nx_train_to_test = x_train[5001:]","metadata":{"execution":{"iopub.status.busy":"2023-06-10T21:38:19.729527Z","iopub.execute_input":"2023-06-10T21:38:19.729926Z","iopub.status.idle":"2023-06-10T21:38:19.738103Z","shell.execute_reply.started":"2023-06-10T21:38:19.729896Z","shell.execute_reply":"2023-06-10T21:38:19.737066Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from keras.layers import Dropout\nfrom keras.regularizers import l2\ndef define_architecture(input_shape, output_shape):\n    inputs = keras.Input(shape=input_shape)  \n    x = layers.Conv2D(filters=64, kernel_size=3, activation=\"relu\", kernel_regularizer=l2(0.01))(inputs)\n    x = layers.MaxPool2D(pool_size=2)(x)\n    x = layers.Conv2D(filters=32, kernel_size=3, activation=\"relu\")(x)\n    x = layers.MaxPool2D(pool_size=2)(x)\n    x = layers.Conv2D(filters=16, kernel_size=3, activation=\"relu\")(x)\n    x = layers.BatchNormalization()(x)\n    x = Dropout(0.5)(x)\n    x = layers.Flatten()(x)\n    outputs = layers.Dense(output_shape, activation=\"softmax\")(x)\n    return keras.Model(inputs=inputs, outputs=outputs)\n\n\nclass RedeNeural(BaseEstimator, ClassifierMixin):\n  def __init__(self, epochs=5, batch_size=64):\n    self.epochs = epochs\n    self.batch_size = batch_size\n\n  def fit(self, X, y):\n    self.labels, ids = np.unique(y, return_inverse=True)\n    yhot = keras.utils.to_categorical(ids)\n    self.model = define_architecture(X.shape[1:], yhot.shape[1])\n    self.model.compile(optimizer=\"rmsprop\",\n                       loss=\"categorical_crossentropy\",\n                       metrics=[\"accuracy\"])\n    self.model.fit(X, yhot, epochs=self.epochs, batch_size=self.batch_size)\n    return self\n\n  def predict(self, X, y=None):\n    probabilities = self.model.predict(X)\n    ypred = self.labels[np.argmax(probabilities, axis=1)]\n    return ypred\n\nclass DividePor255(BaseEstimator, TransformerMixin):\n    def fit(self, X, y):\n        return self\n    def transform(self, X, y=None):\n        return np.array(X, dtype=\"float32\") / 255\n\nclass MudaShape(BaseEstimator, TransformerMixin):\n    def fit(self, X, y):\n        return self\n    def transform(self, X, y=None):\n        return X.reshape((-1,224,224,3))\n\nmodelo = Pipeline([\n    (\"scaler\", DividePor255()), \n    (\"reshape\", MudaShape()), \n    (\"ann\", RedeNeural())\n])\nmodelo.fit(x_train_splited, y_train_splited)\n\nypred = modelo.predict(x_train_to_test)\naccuracy_score(np.array(y_train_to_test), ypred)","metadata":{"execution":{"iopub.status.busy":"2023-06-10T21:40:30.609473Z","iopub.execute_input":"2023-06-10T21:40:30.609838Z","iopub.status.idle":"2023-06-10T22:08:13.412072Z","shell.execute_reply.started":"2023-06-10T21:40:30.609807Z","shell.execute_reply":"2023-06-10T22:08:13.410385Z"},"trusted":true},"execution_count":null,"outputs":[]}]}